|
Post by babaoriley on Mar 28, 2014 8:22:26 GMT -5
Read pages 181 & 182 carefully. Some AF quote material located there. Thoughts by those who understand the science and trials?
|
|
|
Post by mdcenter61 on Mar 28, 2014 8:29:42 GMT -5
Even the way things are worded seems to assume this is moving forward toward an APPROVAL. Darn glad you are aboard - any thoughts on some of the negative aspects on those pages 181 and 182? I'm a CPA nerd and can't decipher much of it but some of that language was concerning.
|
|
|
Post by esstan2001 on Mar 28, 2014 8:40:37 GMT -5
OK mdcenter, don't get too darn glad; I was reading the document prepared by Mannkind... no wonder it sounded so good.
Now on to the FDA version-
|
|
|
Post by slushy on Mar 28, 2014 8:42:15 GMT -5
Here's the conclusion they draw from 181 and 182:
"In conclusion, treatment with TI using Gen2 inhaler was shown to be effective in lowering HbA1c when compared with placebo in the T2DM trial. Based on the protocol-defined non-inferiority margin (0.4%), treatment with TI using Gen2 inhaler was also non-inferior to insulin aspart in lowering HbA1c in the T1DM trial based on the primary analysis. However, because of missing data, the robustness of this analysis is an issue. Since there was only one confirmatory study submitted for the indication of type 1 diabetes mellitus, this makes drawing a solid conclusion regarding efficacy for this type of diabetes mellitus problematic. The final conclusions for approval of the drug/device should also take the comparability of TI and insulin aspart doses as well as safety factors such as hypoglycemia and lung function into consideration."
My take is that it meets the primary end points, but they claim their is missing data. Can someone explain how bad this would be for us? Is it just as easy as supplying more data, or is this something much bigger? Because to me it seems like they're saying we're good to go except for this one glaring issue...
|
|
|
Post by dreggy on Mar 28, 2014 8:54:12 GMT -5
Looks to me like the FDA statisticians have a problem with the trial: "one confirmatory study submitted for the indication of type 1 diabetes mellitus, this makes drawing a solid conclusion regarding efficacy for this type of diabetes mellitus problematic."
They like to see large numbers of comparators, or multiple studies to derive their p values. This looks to me like they want more. In the discussion during the AdCom, the experts other than the stats guys may not have a problem with this, depends on how vocal the stats guy gets.
|
|
|
Post by BD on Mar 28, 2014 8:57:40 GMT -5
From FlyOnTheWall:
09:44 EDT MNKD theflyonthewall.com: FDA reviewer sees issue with limited data for MannKind Afrezza In documents ahead of a FDA panel meeting regarding MannKind's Afrezza inhaler, an FDA reviewer wrote treatment with T1 using the Gen2 inhaler was shown to be effective in lowering HbA1c compared to placebo in a trial. However, because of missing data, the "robustness of this analysis is an issue," the reveiwer wrote, stating that drawing conclusions in Type 1 diabetes is “problematic.” In the document, FDA staff also noted an "imbalance in cases of lung cancers” seen with Afrezza. Reference Link:
|
|
|
Post by BD on Mar 28, 2014 8:58:28 GMT -5
09:46 EDT MNKD theflyonthewall.com: MannKind briefing docs suggest 'uphill battle,' says Piper Jaffray Piper Jaffray believes the FDA panel's briefing documents on Afrezza suggest an "uphill battle" for MannKind into Tuesday's approval meeting. Piper thinks the Type I diabetes indication has little chance of gaining a positive recommendation while the Type II indication is harder to predict. The firm says it is more cautious following the documents given concerns on safety relative to modest efficacy and reiterates a Neutral rating on MannKind. :theflyonthewall
|
|
|
Post by liane on Mar 28, 2014 9:02:09 GMT -5
|
|
|
Post by esstan2001 on Mar 28, 2014 9:27:52 GMT -5
regarding the FDA conclusions...
I think in the AdCom for T1DM there will be a lot of discussion on the data regarding GenII reducing A1C less than IA (insulin aspart- the comparator). I also (in my non expert opinion) do not believe this will b ea big enough issue to derail T1 approval; look at pg 173 Table 5: it is curious, but for men there is statistically no difference between GenII and IA. But there is a 0.4 improvement for women.
Go Figure.
T2DM data seems solid. T1DM data has this question. So what; label it for men only? So here is where it is important for all the other benefits of Afrezza to get weighed properly in the AdCom discussions (reduced excursions, faster acting & clearing, etc)
still reading
|
|
|
Post by harshal1981 on Mar 28, 2014 9:49:28 GMT -5
My short feedback is that FDA doesn't like the at all the fact that in T1, Afrezza met statistical limits with such narrow margin and the fact that it was numerically inferior. They are drawing parallel with study 009 which narrowly missed the statistical limit. They are "wondering" if aspart arm was titrated as frequently as Afrezza Gen2 arm, there was a possibility that even 171 study could have failed.
I DO NOT LIKE WAHT I READ THERE.
175 they are on board.
Safety is non-issue. They are ready to accept pulmonary safety data.
The next big concern is dosing conversion. They do not agree with MNKD approach. PD profile data inadequate to estimate dosing.
More to follow.
|
|
|
Post by liane on Mar 28, 2014 10:06:06 GMT -5
regarding the FDA conclusions... I think in the AdCom for T1DM there will be a lot of discussion on the data regarding GenII reducing A1C less than IA (insulin aspart- the comparator). I also (in my non expert opinion) do not believe this will b ea big enough issue to derail T1 approval; look at pg 173 Table 5: it is curious, but for men there is statistically no difference between GenII and IA. But there is a 0.4 improvement for women. Go Figure. T2DM data seems solid. T1DM data has this question. So what; label it for men only? So here is where it is important for all the other benefits of Afrezza to get weighed properly in the AdCom discussions (reduced excursions, faster acting & clearing, etc) still reading esstan, Welcome! I think you and I are homing in on the same area. Here are the reviewer's comments: Pg 81/248
Reviewer’s comment: The Afrezza TIGen2 group’s use of more basal insulin compared with the insulin aspart group is further evidence that Afrezza TI Gen2 is less effective than insulin aspart, at least as a substitute for prandial subcutaneously injected insulin, i.e . the manner in which Afrezza TI was studied in this trial. The higher basal insulin use the Afrezza TI group is also substantiated by the finding of a lower FPG in the Afrezza TI group at Week 24 compared with the insulin aspart group (see Secondary Endpoints)
Pg 83/248
Reviewer’s comment: It appears that the Afrezza TI groups underwent substantial increases in dose over the study period (by design primarily in the first 12 weeks of the study) whereas the insulin aspart group had a similar dose from start to end of the randomized treatment phase. This finding makes it appear that virtually no titration occurred in the aspart arm.
Nonetheless, it appears both groups were inadequately titrated to reach glycemic goals; the Afrezza TI Gen2 titration algorithm allowed for an increase of 10 U per week, which theoretically would allow for an increase of 120 U over the 12 week prandial insulin titration period. Why the average daily dose only increased by 30 U over the 24 week randomized study period (i.e. mean of 85 U to 115 U) is unclear.
Other observations from these data include: Assuming the stated conversion factor for Gen2 (4 IU aspart = 10 U Afrezza TI) the Gen2 group was using more prandial insulin than the insulin aspart group at Week 24 (115 U Afrezza TI Gen2 is roughly equivalent to 46 IU of rapid acting insulin analog) and overall (103 U Afrezza TI Gen2 is roughly equivalent to 41 IU of rapid acting insulin analog).
It is notable that the aspart group experienced an improvement in HbA1c from Baseline to Week 24 of -0.40% with virtually no increase in the average dose of prandial insulin. Given that the basal insulin optimization phase was only 4 weeks in duration, the effect of basal insulin titration would not be expected to be fully reflected in the Baseline HbA1c. Therefore, the improvement in HbA1c from Baseline to Week 24 was likely driven, in part, by previous titration of basal insulin. Consequently, it is difficult to determine whether there was a reasonable contribution of prandial insulin to the improvement in HbA1c in either treatment arm.
Taking these observations together, it is not clear how to interpret the results of this non-inferiority study. It is concerning that if the insulin aspart group had been titrated more effectively, differences in efficacy between Afrezza TI and insulin aspart might have been greater and the non-inferiority margin may not have been excluded. It is worth reiterating that in a non-inferiority trial design we are basing the efficacy determination on the assumption that the comparator contributed to the effect, and that the within-trial comparator effect size was similar to the historical effect size for trials similarly designed. In T1DM insulin trials this is doubly challenging because two active insulins (basal + prandial) are each contributing to the overall effect.
I think this will get at the whole discussion of how much the basal will do and how much the prandial will do; and whether A1c is the ideal measure. The reviewer was very objective, but since TI is a novel entity, the finesse of titrating it and the basal is not fully appreciated in this study. TI is effective - it is effective over a more narrow window of time. Whether it is quite as "strong" as aspart is debatable. It's a different PK curve - almost comparing apples and oranges.
Bottom line that Adcom and FDA will need to answer - is it safe? and is it effective? Everything else can be worked out over time as it is used in the real world.
|
|
|
Post by mannmade on Mar 28, 2014 10:33:44 GMT -5
So Liane for us non-science folk what does this mean to approval?
|
|
|
Post by esstan2001 on Mar 28, 2014 10:49:55 GMT -5
liane- thank you.
pg 161 corroborates; the statistician has listed under 'other findings and issues'
There may be an issue on whether insulin aspart was optimally given in the control arm.
I think it will come down to how much weighting the other positive Pharmacokinetic and use factors demonstrated in the balance of the other trials will get..
T1DM.... tricky to call T2DM- I think this one happens.
(IMO)
|
|
|
Post by liane on Mar 28, 2014 10:52:53 GMT -5
T1DM.... tricky to call T2DM- I think this one happens. (IMO) And we all know - T2 is the lion's share (>90%) of the market.
|
|
|
Post by noonen on Mar 28, 2014 11:00:29 GMT -5
|
|